Hi,

On 22 September 2016 at 20:53, Felipe Balbi <ba...@kernel.org> wrote:
>
>
> Hi,
>
> Baolin Wang <baolin.w...@linaro.org> writes:
>>>>  static const struct usb_gadget_driver configfs_driver_template = {
>>>>       .bind           = configfs_composite_bind,
>>>>       .unbind         = configfs_composite_unbind,
>>>>
>>>> +#ifdef CONFIG_USB_CONFIGFS_UEVENT
>>>> +     .setup          = configfs_setup,
>>>> +     .reset          = configfs_disconnect,
>>>> +     .disconnect     = configfs_disconnect,
>>>> +#else
>>>>       .setup          = composite_setup,
>>>>       .reset          = composite_disconnect,
>>>>       .disconnect     = composite_disconnect,
>>>> +#endif
>
> nope, this is quite wrong.
>
>>>> @@ -1453,6 +1556,10 @@ static struct config_group *gadgets_make(
>>>>       gi->composite.gadget_driver.function = kstrdup(name, GFP_KERNEL);
>>>>       gi->composite.name = gi->composite.gadget_driver.function;
>>>>
>>>> +#ifdef CONFIG_USB_CONFIGFS_UEVENT
>>>> +     INIT_WORK(&gi->work, configfs_work);
>>>> +#endif
>>>
>>> This is just way too ugly, please make it so there are no #ifdefs in the
>>> .c files.
>>>
>>> Or, as others said, why is this a build option at all, why would you not
>>> always want this enabled if you are relying on it all of the time?
>>
>> Sometimes userspace does not need the notification, it is not all the
>> time. Anyway I will remove the macro if you still insist on that.
>
> what's wrong with the sysfs we already have for this?

If Android system userspace can support udc-core's uevents like Badhri
said, I am fine with that.

-- 
Baolin.wang
Best Regards

Reply via email to